chuk-puzzles-gym 0.9__tar.gz → 0.10.1__tar.gz

{chuk_puzzles_gym-0.9/src/chuk_puzzles_gym.egg-info → chuk_puzzles_gym-0.10.1}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: chuk-puzzles-gym
-Version: 0.9
-Summary: Multi-game puzzle gym for LLM training and benchmarking - 24 constraint puzzles with synthetic data generation
+Version: 0.10.1
+Summary: Multi-game puzzle gym for LLM training and benchmarking - 30 constraint puzzles with synthetic data generation
 Author: Chris Hay
 License: MIT
 Project-URL: Homepage, https://github.com/chrishayuk/chuk-puzzles-gym
@@ -39,7 +39,7 @@ Requires-Dist: ipython>=8.14.0; extra == "dev"
 [![Pydantic v2](https://img.shields.io/badge/pydantic-v2-purple.svg)](https://docs.pydantic.dev/)
 [![Type Checked](https://img.shields.io/badge/type%20checked-mypy-blue.svg)](http://mypy-lang.org/)
 
-A **multi-game puzzle gym** for **LLM training and benchmarking**, hosting 24 different logic puzzle types with synthetic data generation. Built using [chuk-gym-core](https://github.com/chrishayuk/chuk-gym-core) and [chuk-protocol-server](https://github.com/chrishayuk/chuk-protocol-server).
+A **multi-game puzzle gym** for **LLM training and benchmarking**, hosting 30 different logic puzzle types with synthetic data generation. Built using [chuk-gym-core](https://github.com/chrishayuk/chuk-gym-core) and [chuk-protocol-server](https://github.com/chrishayuk/chuk-protocol-server).
 
 **Perfect for:**
 - 🤖 **LLM Agent Testing** - Benchmark reasoning capabilities across constraint types
@@ -82,12 +82,13 @@ Once connected, type `help` to see available games, or `sudoku easy` to start pl
 
 ## Features
 
-- **24 Puzzle Games** with three difficulty levels each (easy, medium, hard)
+- **30 Puzzle Games** with three difficulty levels each (easy, medium, hard)
   - **7 Classic Logic Puzzles** - Sudoku, KenKen, Kakuro, Binary, Futoshiki, Nonogram, Logic Grid
   - **7 Advanced CP-SAT Puzzles** - Killer Sudoku, Lights Out, Mastermind, Slitherlink, Bridges, Hitori, Shikaku
   - **5 Specialized Constraint Puzzles** - Hidato, Tents and Trees, Fillomino, Star Battle, Sokoban
   - **2 Optimization Challenges** - Knapsack, Task Scheduler
   - **3 Advanced Reasoning Puzzles** - Nurikabe, Einstein's Puzzle, Minesweeper
+  - **6 Combinatorial & Search Puzzles** - Skyscrapers, N-Queens, Numberlink, Graph Coloring, Cryptarithmetic, Rush Hour
 - **Agent-Friendly Mode** - Structured output with clear markers for AI agents and tools
   - Enable with `mode agent` command
   - Machine-parseable grid format with clear start/end markers
@@ -113,7 +114,7 @@ Once connected, type `help` to see available games, or `sudoku easy` to start pl
 - **Clean ASCII art grids** - perfectly aligned for easy parsing
 - **Deterministic seeding** - Replay any puzzle with the same seed
 - **Gymnasium-compatible RL Environment** (`PuzzleEnv`) for training agents
-- **Comprehensive test suite** (1067 tests, 94% coverage)
+- **Comprehensive test suite** (1323 tests, 94% coverage)
 - **Modern Python best practices:**
   - **Pydantic v2 native** - All models use ConfigDict for type safety
   - **Async native** - Full async/await support throughout
@@ -173,6 +174,17 @@ Once connected, type `help` to see available games, or `sudoku easy` to start pl
 | **Einstein's Puzzle** | 5 houses × 5 attributes | Multi-attribute deduction + Logic chains | ✅ Complete |
 | **Minesweeper** | 6×6 to 10×10 | Probabilistic reasoning + Safe deduction | ✅ Complete |
 
+### Combinatorial & Search Puzzles
+
+| Game | Grid Size | Constraint Types | Status |
+|------|-----------|------------------|--------|
+| **Skyscrapers** | 4×4 to 6×6 | Latin square + Visibility clues from 4 borders | ✅ Complete |
+| **N-Queens** | 6×6 to 12×12 | Placement + Row/Column/Diagonal attack avoidance | ✅ Complete |
+| **Numberlink** | 5×5 to 9×9 | Path connectivity + Non-crossing + Space filling | ✅ Complete |
+| **Graph Coloring** | 6-15 nodes | Graph coloring + Inequality + Global constraint | ✅ Complete |
+| **Cryptarithmetic** | 3-5 digit words | Arithmetic + AllDifferent + Carry propagation | ✅ Complete |
+| **Rush Hour** | 6×6 | Sequential planning + Spatial blocking + Search | ✅ Complete |
+
 ## Solver Profiles & Business Mapping
 
 Each game includes metadata for **constraint types**, **business analogies**, and **complexity profiles**, making it easy to:
@@ -216,7 +228,13 @@ resource_games = [
 | **Global Loop** | Slitherlink | Circuit design, Path finding |
 | **Boolean SAT** | Lights Out | Feature dependencies, Toggle systems |
 | **Cage Sums** | Killer Sudoku, Kakuro | Team budgets, Grouped constraints |
-| **AllDifferent** | Sudoku, KenKen | Resource uniqueness, Assignment problems |
+| **AllDifferent** | Sudoku, KenKen, Skyscrapers | Resource uniqueness, Assignment problems |
+| **Visibility/Ordering** | Skyscrapers | Priority ranking, Stack-based processing |
+| **Attack Avoidance** | N-Queens, Star Battle | Non-conflicting resource placement |
+| **Path Connectivity** | Numberlink, Nurikabe | Network routing, Cable layout |
+| **Graph Coloring** | Graph Coloring | Frequency assignment, Register allocation, Scheduling |
+| **Arithmetic Deduction** | Cryptarithmetic, KenKen | Code breaking, Constraint propagation |
+| **Sequential Planning** | Rush Hour, Sokoban | Logistics planning, Deadlock resolution |
 
 ## Quick Start
 
@@ -397,6 +415,14 @@ ADVANCED REASONING PUZZLES:
  23) Einstein's Puzzle - Who owns the fish? Multi-attribute deduction
  24) Minesweeper     - Find all mines using logical deduction
 
+COMBINATORIAL & SEARCH PUZZLES:
+ 25) Skyscrapers     - Latin square with visibility clues from borders
+ 26) N-Queens        - Place queens with no row/column/diagonal conflicts
+ 27) Numberlink      - Connect pairs with non-crossing paths filling the grid
+ 28) Graph Coloring  - Color nodes so no adjacent pair shares a color
+ 29) Cryptarithmetic - Assign digits to letters to satisfy an equation
+ 30) Rush Hour       - Slide vehicles to free the target car to the exit
+
 Commands:
   <number>  - Select game by number
   <name>    - Select game by name (e.g., 'sudoku')
@@ -453,7 +479,7 @@ The project includes a **Gymnasium-compatible environment** for training reinfor
 ```python
 from chuk_puzzles_gym.gym_env import PuzzleEnv
 
-# Create environment for any of the 24 games
+# Create environment for any of the 30 games
 env = PuzzleEnv("sudoku", difficulty="easy", seed=42)
 
 # Reset to start a new episode
@@ -472,7 +498,7 @@ games = PuzzleEnv.available_games()
 
 ### Features
 
-- **All 24 games** accessible through unified API
+- **All 30 games** accessible through unified API
 - **Configurable rewards** for correct moves, invalid attempts, completion bonuses
 - **Hint system** with optional budget limits
 - **Solver-free mode** for pure reasoning benchmarks
@@ -586,7 +612,7 @@ Generate synthetic puzzle datasets for training and benchmarking LLMs and constr
 ### CLI Usage
 
 ```bash
-# Generate 100 puzzles per game/difficulty for all 24 games
+# Generate 100 puzzles per game/difficulty for all 30 games
 chuk-puzzles-export -o puzzles.jsonl
 
 # Specific games only
@@ -744,6 +770,9 @@ When `include_trace=True` (default), each problem includes step-by-step solution
 | Hitori | `duplicate_elimination` |
 | Bridges | `connectivity_constraint` |
 | Slitherlink | `loop_constraint` |
+| Graph Coloring | `graph_coloring_constraint` |
+| Cryptarithmetic | `arithmetic_constraint` |
+| Rush Hour | `sequential_planning` |
 | Others | `constraint_propagation` |
 
 ### Example: Generate Training Data
@@ -772,9 +801,9 @@ With default settings (`-n 100` per game/difficulty):
 
 | Configuration | Problems Generated |
 |--------------|-------------------|
-| All games, all difficulties | 24 games × 3 difficulties × 100 = 7,200 |
+| All games, all difficulties | 30 games × 3 difficulties × 100 = 9,000 |
 | Single game, all difficulties | 1 × 3 × 100 = 300 |
-| All games, single difficulty | 24 × 1 × 100 = 2,400 |
+| All games, single difficulty | 30 × 1 × 100 = 3,000 |
 
 ### Integration with chuk-gym-core
 
@@ -983,7 +1012,7 @@ pip install -e ".[dev]"
 
 ### Testing
 
-The project has comprehensive test coverage (94%, 1067 tests):
+The project has comprehensive test coverage (94%, 1323 tests):
 
 ```bash
 # Run all tests
@@ -1034,7 +1063,7 @@ The project follows modern Python best practices with a **9.8/10 compliance scor
 - ✅ **Test Coverage** (9.5/10) - 94% overall, most files ≥90%
 
 #### Quality Metrics
-- **1067 tests** - All passing ✅
+- **1323 tests** - All passing ✅
 - **94% coverage** - Exceeds 90% threshold ✅
 - **Zero linting errors** - Clean codebase ✅
 - **Full type safety** - MyPy passes ✅
@@ -1181,7 +1210,7 @@ chuk-puzzles-gym/
 │           │   ├── __init__.py
 │           │   ├── game.py
 │           │   └── config.py
-│           └── ... (24 games total)
+│           └── ... (30 games total)
 ├── tests/
 │   ├── test_puzzle_game.py       # Base class tests
 │   ├── test_deterministic_seeding.py  # Seeding tests
@@ -1191,6 +1220,12 @@ chuk-puzzles-gym/
 ├── examples/
 │   ├── simple_client.py          # Telnet client example
 │   ├── websocket_client.py       # WebSocket client example
+│   ├── example_skyscrapers.py    # Skyscrapers game logic demo
+│   ├── example_nqueens.py        # N-Queens game logic demo
+│   ├── example_numberlink.py     # Numberlink game logic demo
+│   ├── example_graph_coloring.py # Graph Coloring game logic demo
+│   ├── example_cryptarithmetic.py# Cryptarithmetic game logic demo
+│   ├── example_rush_hour.py      # Rush Hour game logic demo
 │   └── README.md                 # Example usage guide
 ├── .github/workflows/            # CI/CD workflows
 ├── pyproject.toml                # Modern Python project config
@@ -1203,14 +1238,15 @@ chuk-puzzles-gym/
 
 ### Key Statistics
 
-- **Test Coverage**: 94% overall (1067 tests, all passing)
+- **Test Coverage**: 94% overall (1323 tests, all passing)
 - **Code Quality Score**: 9.8/10 (near perfect compliance)
-- **Games Implemented**: 24 complete puzzle types
+- **Games Implemented**: 30 complete puzzle types
   - 7 Classic Logic Puzzles
   - 7 Advanced CP-SAT Puzzles
   - 5 Specialized Constraint Puzzles
   - 2 Optimization Challenges
   - 3 Advanced Reasoning Puzzles
+  - 6 Combinatorial & Search Puzzles
 - **Supported Transports**: 4 (Telnet, TCP, WebSocket, WS-Telnet)
 - **Agent-Friendly Mode**: Structured output for AI tools
 - **Gymnasium API**: RL-compatible environment for all games
@@ -1243,7 +1279,7 @@ Test the generality of constraint solvers (like MCP solvers):
 
 Learn about constraint satisfaction problems:
 
-- **24 different puzzle types** demonstrating various constraint types:
+- **30 different puzzle types** demonstrating various constraint types:
   - AllDifferent constraints (Sudoku, KenKen, Futoshiki)
   - Arithmetic constraints (KenKen, Kakuro, Killer Sudoku)
   - Boolean/SAT constraints (Lights Out, Binary Puzzle)
@@ -1253,9 +1289,14 @@ Learn about constraint satisfaction problems:
   - Temporal reasoning (Task Scheduler)
   - Connectivity constraints (Nurikabe, Slitherlink)
   - Probabilistic reasoning (Minesweeper)
-  - And more!
+  - Graph coloring (Graph Coloring)
+  - Arithmetic deduction (Cryptarithmetic)
+  - Sequential planning (Rush Hour)
+  - Visibility constraints (Skyscrapers)
+  - Attack avoidance (N-Queens)
+  - Path connectivity (Numberlink)
 - **Well-documented code** showing puzzle generation algorithms
-- **Comprehensive tests** (1067 tests, 94% coverage) demonstrating validation
+- **Comprehensive tests** (1323 tests, 94% coverage) demonstrating validation
 - **Deterministic seeding** - Reproduce any puzzle for debugging/testing
 - **Production-ready** - 9.8/10 code quality score
 - **Type-safe** - Full Pydantic v2 and MyPy compliance

{chuk_puzzles_gym-0.9 → chuk_puzzles_gym-0.10.1}/README.md

@@ -8,7 +8,7 @@
 [![Pydantic v2](https://img.shields.io/badge/pydantic-v2-purple.svg)](https://docs.pydantic.dev/)
 [![Type Checked](https://img.shields.io/badge/type%20checked-mypy-blue.svg)](http://mypy-lang.org/)
 
-A **multi-game puzzle gym** for **LLM training and benchmarking**, hosting 24 different logic puzzle types with synthetic data generation. Built using [chuk-gym-core](https://github.com/chrishayuk/chuk-gym-core) and [chuk-protocol-server](https://github.com/chrishayuk/chuk-protocol-server).
+A **multi-game puzzle gym** for **LLM training and benchmarking**, hosting 30 different logic puzzle types with synthetic data generation. Built using [chuk-gym-core](https://github.com/chrishayuk/chuk-gym-core) and [chuk-protocol-server](https://github.com/chrishayuk/chuk-protocol-server).
 
 **Perfect for:**
 - 🤖 **LLM Agent Testing** - Benchmark reasoning capabilities across constraint types
@@ -51,12 +51,13 @@ Once connected, type `help` to see available games, or `sudoku easy` to start pl
 
 ## Features
 
-- **24 Puzzle Games** with three difficulty levels each (easy, medium, hard)
+- **30 Puzzle Games** with three difficulty levels each (easy, medium, hard)
   - **7 Classic Logic Puzzles** - Sudoku, KenKen, Kakuro, Binary, Futoshiki, Nonogram, Logic Grid
   - **7 Advanced CP-SAT Puzzles** - Killer Sudoku, Lights Out, Mastermind, Slitherlink, Bridges, Hitori, Shikaku
   - **5 Specialized Constraint Puzzles** - Hidato, Tents and Trees, Fillomino, Star Battle, Sokoban
   - **2 Optimization Challenges** - Knapsack, Task Scheduler
   - **3 Advanced Reasoning Puzzles** - Nurikabe, Einstein's Puzzle, Minesweeper
+  - **6 Combinatorial & Search Puzzles** - Skyscrapers, N-Queens, Numberlink, Graph Coloring, Cryptarithmetic, Rush Hour
 - **Agent-Friendly Mode** - Structured output with clear markers for AI agents and tools
   - Enable with `mode agent` command
   - Machine-parseable grid format with clear start/end markers
@@ -82,7 +83,7 @@ Once connected, type `help` to see available games, or `sudoku easy` to start pl
 - **Clean ASCII art grids** - perfectly aligned for easy parsing
 - **Deterministic seeding** - Replay any puzzle with the same seed
 - **Gymnasium-compatible RL Environment** (`PuzzleEnv`) for training agents
-- **Comprehensive test suite** (1067 tests, 94% coverage)
+- **Comprehensive test suite** (1323 tests, 94% coverage)
 - **Modern Python best practices:**
   - **Pydantic v2 native** - All models use ConfigDict for type safety
   - **Async native** - Full async/await support throughout
@@ -142,6 +143,17 @@ Once connected, type `help` to see available games, or `sudoku easy` to start pl
 | **Einstein's Puzzle** | 5 houses × 5 attributes | Multi-attribute deduction + Logic chains | ✅ Complete |
 | **Minesweeper** | 6×6 to 10×10 | Probabilistic reasoning + Safe deduction | ✅ Complete |
 
+### Combinatorial & Search Puzzles
+
+| Game | Grid Size | Constraint Types | Status |
+|------|-----------|------------------|--------|
+| **Skyscrapers** | 4×4 to 6×6 | Latin square + Visibility clues from 4 borders | ✅ Complete |
+| **N-Queens** | 6×6 to 12×12 | Placement + Row/Column/Diagonal attack avoidance | ✅ Complete |
+| **Numberlink** | 5×5 to 9×9 | Path connectivity + Non-crossing + Space filling | ✅ Complete |
+| **Graph Coloring** | 6-15 nodes | Graph coloring + Inequality + Global constraint | ✅ Complete |
+| **Cryptarithmetic** | 3-5 digit words | Arithmetic + AllDifferent + Carry propagation | ✅ Complete |
+| **Rush Hour** | 6×6 | Sequential planning + Spatial blocking + Search | ✅ Complete |
+
 ## Solver Profiles & Business Mapping
 
 Each game includes metadata for **constraint types**, **business analogies**, and **complexity profiles**, making it easy to:
@@ -185,7 +197,13 @@ resource_games = [
 | **Global Loop** | Slitherlink | Circuit design, Path finding |
 | **Boolean SAT** | Lights Out | Feature dependencies, Toggle systems |
 | **Cage Sums** | Killer Sudoku, Kakuro | Team budgets, Grouped constraints |
-| **AllDifferent** | Sudoku, KenKen | Resource uniqueness, Assignment problems |
+| **AllDifferent** | Sudoku, KenKen, Skyscrapers | Resource uniqueness, Assignment problems |
+| **Visibility/Ordering** | Skyscrapers | Priority ranking, Stack-based processing |
+| **Attack Avoidance** | N-Queens, Star Battle | Non-conflicting resource placement |
+| **Path Connectivity** | Numberlink, Nurikabe | Network routing, Cable layout |
+| **Graph Coloring** | Graph Coloring | Frequency assignment, Register allocation, Scheduling |
+| **Arithmetic Deduction** | Cryptarithmetic, KenKen | Code breaking, Constraint propagation |
+| **Sequential Planning** | Rush Hour, Sokoban | Logistics planning, Deadlock resolution |
 
 ## Quick Start
 
@@ -366,6 +384,14 @@ ADVANCED REASONING PUZZLES:
  23) Einstein's Puzzle - Who owns the fish? Multi-attribute deduction
  24) Minesweeper     - Find all mines using logical deduction
 
+COMBINATORIAL & SEARCH PUZZLES:
+ 25) Skyscrapers     - Latin square with visibility clues from borders
+ 26) N-Queens        - Place queens with no row/column/diagonal conflicts
+ 27) Numberlink      - Connect pairs with non-crossing paths filling the grid
+ 28) Graph Coloring  - Color nodes so no adjacent pair shares a color
+ 29) Cryptarithmetic - Assign digits to letters to satisfy an equation
+ 30) Rush Hour       - Slide vehicles to free the target car to the exit
+
 Commands:
   <number>  - Select game by number
   <name>    - Select game by name (e.g., 'sudoku')
@@ -422,7 +448,7 @@ The project includes a **Gymnasium-compatible environment** for training reinfor
 ```python
 from chuk_puzzles_gym.gym_env import PuzzleEnv
 
-# Create environment for any of the 24 games
+# Create environment for any of the 30 games
 env = PuzzleEnv("sudoku", difficulty="easy", seed=42)
 
 # Reset to start a new episode
@@ -441,7 +467,7 @@ games = PuzzleEnv.available_games()
 
 ### Features
 
-- **All 24 games** accessible through unified API
+- **All 30 games** accessible through unified API
 - **Configurable rewards** for correct moves, invalid attempts, completion bonuses
 - **Hint system** with optional budget limits
 - **Solver-free mode** for pure reasoning benchmarks
@@ -555,7 +581,7 @@ Generate synthetic puzzle datasets for training and benchmarking LLMs and constr
 ### CLI Usage
 
 ```bash
-# Generate 100 puzzles per game/difficulty for all 24 games
+# Generate 100 puzzles per game/difficulty for all 30 games
 chuk-puzzles-export -o puzzles.jsonl
 
 # Specific games only
@@ -713,6 +739,9 @@ When `include_trace=True` (default), each problem includes step-by-step solution
 | Hitori | `duplicate_elimination` |
 | Bridges | `connectivity_constraint` |
 | Slitherlink | `loop_constraint` |
+| Graph Coloring | `graph_coloring_constraint` |
+| Cryptarithmetic | `arithmetic_constraint` |
+| Rush Hour | `sequential_planning` |
 | Others | `constraint_propagation` |
 
 ### Example: Generate Training Data
@@ -741,9 +770,9 @@ With default settings (`-n 100` per game/difficulty):
 
 | Configuration | Problems Generated |
 |--------------|-------------------|
-| All games, all difficulties | 24 games × 3 difficulties × 100 = 7,200 |
+| All games, all difficulties | 30 games × 3 difficulties × 100 = 9,000 |
 | Single game, all difficulties | 1 × 3 × 100 = 300 |
-| All games, single difficulty | 24 × 1 × 100 = 2,400 |
+| All games, single difficulty | 30 × 1 × 100 = 3,000 |
 
 ### Integration with chuk-gym-core
 
@@ -952,7 +981,7 @@ pip install -e ".[dev]"
 
 ### Testing
 
-The project has comprehensive test coverage (94%, 1067 tests):
+The project has comprehensive test coverage (94%, 1323 tests):
 
 ```bash
 # Run all tests
@@ -1003,7 +1032,7 @@ The project follows modern Python best practices with a **9.8/10 compliance scor
 - ✅ **Test Coverage** (9.5/10) - 94% overall, most files ≥90%
 
 #### Quality Metrics
-- **1067 tests** - All passing ✅
+- **1323 tests** - All passing ✅
 - **94% coverage** - Exceeds 90% threshold ✅
 - **Zero linting errors** - Clean codebase ✅
 - **Full type safety** - MyPy passes ✅
@@ -1150,7 +1179,7 @@ chuk-puzzles-gym/
 │           │   ├── __init__.py
 │           │   ├── game.py
 │           │   └── config.py
-│           └── ... (24 games total)
+│           └── ... (30 games total)
 ├── tests/
 │   ├── test_puzzle_game.py       # Base class tests
 │   ├── test_deterministic_seeding.py  # Seeding tests
@@ -1160,6 +1189,12 @@ chuk-puzzles-gym/
 ├── examples/
 │   ├── simple_client.py          # Telnet client example
 │   ├── websocket_client.py       # WebSocket client example
+│   ├── example_skyscrapers.py    # Skyscrapers game logic demo
+│   ├── example_nqueens.py        # N-Queens game logic demo
+│   ├── example_numberlink.py     # Numberlink game logic demo
+│   ├── example_graph_coloring.py # Graph Coloring game logic demo
+│   ├── example_cryptarithmetic.py# Cryptarithmetic game logic demo
+│   ├── example_rush_hour.py      # Rush Hour game logic demo
 │   └── README.md                 # Example usage guide
 ├── .github/workflows/            # CI/CD workflows
 ├── pyproject.toml                # Modern Python project config
@@ -1172,14 +1207,15 @@ chuk-puzzles-gym/
 
 ### Key Statistics
 
-- **Test Coverage**: 94% overall (1067 tests, all passing)
+- **Test Coverage**: 94% overall (1323 tests, all passing)
 - **Code Quality Score**: 9.8/10 (near perfect compliance)
-- **Games Implemented**: 24 complete puzzle types
+- **Games Implemented**: 30 complete puzzle types
   - 7 Classic Logic Puzzles
   - 7 Advanced CP-SAT Puzzles
   - 5 Specialized Constraint Puzzles
   - 2 Optimization Challenges
   - 3 Advanced Reasoning Puzzles
+  - 6 Combinatorial & Search Puzzles
 - **Supported Transports**: 4 (Telnet, TCP, WebSocket, WS-Telnet)
 - **Agent-Friendly Mode**: Structured output for AI tools
 - **Gymnasium API**: RL-compatible environment for all games
@@ -1212,7 +1248,7 @@ Test the generality of constraint solvers (like MCP solvers):
 
 Learn about constraint satisfaction problems:
 
-- **24 different puzzle types** demonstrating various constraint types:
+- **30 different puzzle types** demonstrating various constraint types:
   - AllDifferent constraints (Sudoku, KenKen, Futoshiki)
   - Arithmetic constraints (KenKen, Kakuro, Killer Sudoku)
   - Boolean/SAT constraints (Lights Out, Binary Puzzle)
@@ -1222,9 +1258,14 @@ Learn about constraint satisfaction problems:
   - Temporal reasoning (Task Scheduler)
   - Connectivity constraints (Nurikabe, Slitherlink)
   - Probabilistic reasoning (Minesweeper)
-  - And more!
+  - Graph coloring (Graph Coloring)
+  - Arithmetic deduction (Cryptarithmetic)
+  - Sequential planning (Rush Hour)
+  - Visibility constraints (Skyscrapers)
+  - Attack avoidance (N-Queens)
+  - Path connectivity (Numberlink)
 - **Well-documented code** showing puzzle generation algorithms
-- **Comprehensive tests** (1067 tests, 94% coverage) demonstrating validation
+- **Comprehensive tests** (1323 tests, 94% coverage) demonstrating validation
 - **Deterministic seeding** - Reproduce any puzzle for debugging/testing
 - **Production-ready** - 9.8/10 code quality score
 - **Type-safe** - Full Pydantic v2 and MyPy compliance

{chuk_puzzles_gym-0.9 → chuk_puzzles_gym-0.10.1}/pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "chuk-puzzles-gym"
-version = "0.9"
-description = "Multi-game puzzle gym for LLM training and benchmarking - 24 constraint puzzles with synthetic data generation"
+version = "0.10.1"
+description = "Multi-game puzzle gym for LLM training and benchmarking - 30 constraint puzzles with synthetic data generation"
 readme = "README.md"
 requires-python = ">=3.11"
 license = {text = "MIT"}

{chuk_puzzles_gym-0.9 → chuk_puzzles_gym-0.10.1}/src/chuk_puzzles_gym/eval.py

@@ -235,6 +235,9 @@ async def _apply_hint(game: PuzzleGame, hint_data: tuple) -> MoveResult:
         "nonogram",
         "hidato",
         "fillomino",
+        "skyscrapers",
+        "n-queens",
+        "numberlink",
     ]:
         if len(hint_data) >= 3:
             row, col, value = hint_data[0], hint_data[1], hint_data[2]
@@ -342,6 +345,24 @@ async def _apply_hint(game: PuzzleGame, hint_data: tuple) -> MoveResult:
             direction = hint_data if isinstance(hint_data, str) else hint_data
             return await game.validate_move(direction)
 
+    # Graph Coloring - hint is (node, color)
+    if game_name in ["graph coloring"]:
+        if len(hint_data) >= 2:
+            node, color = hint_data[0], hint_data[1]
+            return await game.validate_move(node, color)
+
+    # Cryptarithmetic - hint is (letter, digit)
+    if game_name in ["cryptarithmetic"]:
+        if len(hint_data) >= 2:
+            letter, digit = hint_data[0], hint_data[1]
+            return await game.validate_move(letter, digit)
+
+    # Rush Hour - hint is (vehicle_id, direction)
+    if game_name in ["rush hour"]:
+        if len(hint_data) >= 2:
+            vehicle_id, direction = hint_data[0], hint_data[1]
+            return await game.validate_move(vehicle_id, direction)
+
     # Generic fallback - try validate_move with hint args as tuple
     if isinstance(hint_data, tuple) and len(hint_data) >= 2:
         return await game.validate_move(*hint_data)

{chuk_puzzles_gym-0.9 → chuk_puzzles_gym-0.10.1}/src/chuk_puzzles_gym/games/__init__.py

@@ -2,9 +2,11 @@
 
 from .binary import BinaryPuzzleGame
 from .bridges import BridgesGame
+from .cryptarithmetic import CryptarithmeticCommandHandler, CryptarithmeticGame
 from .einstein import EinsteinGame
 from .fillomino import FillominoGame
 from .futoshiki import FutoshikiGame
+from .graph_coloring import GraphColoringCommandHandler, GraphColoringGame
 from .hidato import HidatoGame
 from .hitori import HitoriGame
 from .kakuro import KakuroGame
@@ -16,9 +18,13 @@ from .logic_grid import LogicGridGame
 from .mastermind import MastermindGame
 from .minesweeper import MinesweeperGame
 from .nonogram import NonogramGame
+from .nqueens import NQueensGame
+from .numberlink import NumberlinkGame
 from .nurikabe import NurikabeGame
+from .rush_hour import RushHourCommandHandler, RushHourGame
 from .scheduler import SchedulerGame
 from .shikaku import ShikakuGame
+from .skyscrapers import SkyscrapersGame
 from .slitherlink import SlitherlinkGame
 from .sokoban import SokobanGame
 from .star_battle import StarBattleGame
@@ -57,11 +63,21 @@ AVAILABLE_GAMES = {
     "nurikabe": NurikabeGame,
     "einstein": EinsteinGame,
     "minesweeper": MinesweeperGame,
+    # New Games
+    "skyscrapers": SkyscrapersGame,
+    "nqueens": NQueensGame,
+    "numberlink": NumberlinkGame,
+    "graph_coloring": GraphColoringGame,
+    "cryptarithmetic": CryptarithmeticGame,
+    "rush_hour": RushHourGame,
 }
 
 # Registry of game command handlers (games that have moved command handling out of server)
 GAME_COMMAND_HANDLERS = {
     "sudoku": SudokuCommandHandler,
+    "graph_coloring": GraphColoringCommandHandler,
+    "cryptarithmetic": CryptarithmeticCommandHandler,
+    "rush_hour": RushHourCommandHandler,
 }
 
 __all__ = [
@@ -89,6 +105,12 @@ __all__ = [
     "NurikabeGame",
     "EinsteinGame",
     "MinesweeperGame",
+    "SkyscrapersGame",
+    "NQueensGame",
+    "NumberlinkGame",
+    "GraphColoringGame",
+    "CryptarithmeticGame",
+    "RushHourGame",
     "AVAILABLE_GAMES",
     "GAME_COMMAND_HANDLERS",
 ]

{chuk_puzzles_gym-0.9 → chuk_puzzles_gym-0.10.1}/src/chuk_puzzles_gym/games/binary/game.py

@@ -358,6 +358,8 @@ class BinaryPuzzleGame(PuzzleGame):
         Returns:
             Tuple of (hint_data, hint_message) or None if puzzle is complete
         """
+        if not self.can_use_hint():
+            return None
         empty_cells = [(r, c) for r in range(self.size) for c in range(self.size) if self.grid[r][c] == -1]
         if not empty_cells:
             return None

{chuk_puzzles_gym-0.9 → chuk_puzzles_gym-0.10.1}/src/chuk_puzzles_gym/games/bridges/game.py

@@ -415,6 +415,8 @@ class BridgesGame(PuzzleGame):
 
     async def get_hint(self) -> tuple[Any, str] | None:
         """Get a hint for the next move."""
+        if not self.can_use_hint():
+            return None
         # Find a bridge in the solution that's not yet placed correctly
         for bridge_key, solution_count in self.solution.items():
             current_count = self.bridges.get(bridge_key, 0)

chuk_puzzles_gym-0.10.1/src/chuk_puzzles_gym/games/cryptarithmetic/__init__.py

@@ -0,0 +1,7 @@
+"""Cryptarithmetic puzzle game."""
+
+from .commands import CryptarithmeticCommandHandler
+from .config import CryptarithmeticConfig
+from .game import CryptarithmeticGame
+
+__all__ = ["CryptarithmeticGame", "CryptarithmeticConfig", "CryptarithmeticCommandHandler"]

chuk_puzzles_gym-0.10.1/src/chuk_puzzles_gym/games/cryptarithmetic/commands.py

@@ -0,0 +1,75 @@
+"""Command handler for Cryptarithmetic game."""
+
+from typing import TYPE_CHECKING
+
+from ...models import GameCommand, MoveResult
+from .._base import CommandResult, GameCommandHandler
+
+if TYPE_CHECKING:
+    from .game import CryptarithmeticGame
+
+
+class CryptarithmeticCommandHandler(GameCommandHandler):
+    """Handles commands for Cryptarithmetic game."""
+
+    game: "CryptarithmeticGame"
+
+    @property
+    def supported_commands(self) -> set[GameCommand]:
+        """Return the set of GameCommand enums this handler supports."""
+        return {GameCommand.ASSIGN, GameCommand.UNASSIGN}
+
+    async def handle_command(self, cmd: GameCommand, args: list[str]) -> CommandResult:
+        """Handle a Cryptarithmetic command.
+
+        Args:
+            cmd: The GameCommand enum value
+            args: List of string arguments (already split from input)
+
+        Returns:
+            CommandResult with the move result and display flags
+        """
+        if cmd == GameCommand.ASSIGN:
+            return await self._handle_assign(args)
+        elif cmd == GameCommand.UNASSIGN:
+            return await self._handle_unassign(args)
+        else:
+            return self.error_result(f"Unknown command: {cmd}")
+
+    async def _handle_assign(self, args: list[str]) -> CommandResult:
+        """Handle the ASSIGN command: assign <letter> <digit>."""
+        if len(args) != 2:
+            return CommandResult(
+                result=MoveResult(success=False, message="Usage: assign <letter> <digit>"),
+                should_display=False,
+            )
+
+        letter = args[0].upper()
+        digit = self.parse_int(args[1], "digit")
+
+        if digit is None:
+            return self.error_result("Digit must be an integer (0-9).")
+
+        result = await self.game.validate_move(letter, digit)
+
+        return CommandResult(
+            result=result,
+            should_display=result.success,
+            is_game_over=result.success and self.game.is_complete(),
+        )
+
+    async def _handle_unassign(self, args: list[str]) -> CommandResult:
+        """Handle the UNASSIGN command: unassign <letter>."""
+        if len(args) != 1:
+            return CommandResult(
+                result=MoveResult(success=False, message="Usage: unassign <letter>"),
+                should_display=False,
+            )
+
+        letter = args[0].upper()
+        result = await self.game.validate_move(letter, -1)
+
+        return CommandResult(
+            result=result,
+            should_display=result.success,
+        )

chuk_puzzles_gym-0.10.1/src/chuk_puzzles_gym/games/cryptarithmetic/config.py

@@ -0,0 +1,23 @@
+"""Configuration for Cryptarithmetic puzzle game."""
+
+from pydantic import BaseModel, Field
+
+from ...models import DifficultyLevel
+
+
+class CryptarithmeticConfig(BaseModel):
+    """Configuration for a Cryptarithmetic puzzle."""
+
+    difficulty: DifficultyLevel = Field(default=DifficultyLevel.EASY)
+    max_word_length: int = Field(ge=2, le=6, description="Maximum word length")
+    pre_assigned: int = Field(ge=0, description="Number of pre-assigned letter-digit pairs")
+
+    @classmethod
+    def from_difficulty(cls, difficulty: DifficultyLevel) -> "CryptarithmeticConfig":
+        """Create config from difficulty level."""
+        config_map = {
+            DifficultyLevel.EASY: {"max_word_length": 3, "pre_assigned": 3},
+            DifficultyLevel.MEDIUM: {"max_word_length": 4, "pre_assigned": 2},
+            DifficultyLevel.HARD: {"max_word_length": 5, "pre_assigned": 0},
+        }
+        return cls(difficulty=difficulty, **config_map[difficulty])