---

name: reasoner description: > Advanced reasoning with search strategies (beam search, MCTS). WHEN: Complex problem-solving requiring exploration of multiple solution paths, optimization problems, decision trees, when you need scored/ranked reasoning paths. WHEN NOT: Simple linear reasoning (use sequential_thinking), trivial problems, when branching isn't needed. version: 0.1.0

Reasoner - Advanced Multi-Strategy Reasoning

Core Concept

mcp__plugin_kg_kodegen__reasoner provides sophisticated reasoning with multiple search strategies. Unlike sequential_thinking (simple linear tracking), reasoner uses algorithms like Beam Search and Monte Carlo Tree Search (MCTS) to explore and score multiple solution paths, finding optimal reasoning chains.

Strategies

Key Parameters

Required:

Optional:

Usage Examples

Beam Search (Default)

Copy{
  "thought": "Analyzing possible caching strategies for the API",
  "thought_number": 1,
  "total_thoughts": 4,
  "next_thought_needed": true,
  "strategy_type": "beam_search",
  "beam_width": 3
}

MCTS for Decision Making

Copy{
  "thought": "Evaluating database migration approaches",
  "thought_number": 1,
  "total_thoughts": 3,
  "next_thought_needed": true,
  "strategy_type": "mcts",
  "num_simulations": 100
}

Creative Problem Solving

Copy{
  "thought": "Exploring novel approaches to distributed consensus",
  "thought_number": 1,
  "total_thoughts": 5,
  "next_thought_needed": true,
  "strategy_type": "mcts_002_alpha",
  "num_simulations": 75
}

Detailed Analysis

Copy{
  "thought": "Deep comparison of microservices vs monolithic architecture",
  "thought_number": 1,
  "total_thoughts": 6,
  "next_thought_needed": true,
  "strategy_type": "mcts_002alt_alpha",
  "num_simulations": 50
}

Branching from Parent

Copy{
  "thought": "Alternative approach using event sourcing",
  "thought_number": 3,
  "total_thoughts": 5,
  "next_thought_needed": true,
  "parent_id": "previous-node-uuid"
}

Output Format

Copy{
  "session_id": "uuid-v4",
  "thought": "echoed input",
  "score": 0.85,
  "depth": 2,
  "is_complete": false,
  "next_thought_needed": true,
  "branches": 3,
  "best_path_score": 0.92,
  "strategy": "beam_search",
  "history_length": 5
}

When to Use What

Reasoner vs Sequential Thinking

Remember

• Choose strategy wisely - beam_search for general, MCTS for optimization
• Adjust beam_width - higher = more paths but slower
• num_simulations - more = better MCTS results but slower
• Check scores - output includes quality scores (0.0-1.0)
• Use for complex problems - overkill for simple reasoning
• Prefer sequential_thinking for straightforward step-by-step