Reasoning Patterns

Reasoning capabilities are exposed as reusable pattern implementations rather than prompt-only conventions.

Available patterns

• ProposeCriticPattern - Iterative propose/critic refinement. - Background references: Self-Refine; Reflexion. Conceptual grounding only; behavior is defined by repository contracts and implementation.

• BeamSearchPattern - Generator + evaluator delegate orchestration with max_depth, branch_factor, and beam_width controls. - Background references: Tree of Thoughts. Conceptual grounding only; this implementation uses framework-native step orchestration.

• RAGPattern - Retrieval-augmented reasoning with memory read/write workflow primitives. - Background references: Retrieval-Augmented Generation (RAG). Conceptual grounding only; this pattern composes retrieval and context injection at workflow level.

Beam search output

BeamSearchPattern returns:

{
    "final_output": {
        "best_candidate": {...},
        "best_score": float,
    },
    "details": {
        "explored_nodes": int,
        "frontier_trace": [...],
    },
    "terminated_reason": str,
}

Examples

• examples/patterns/beam_search.py

• examples/patterns/propose_critic.py

• examples/patterns/rag.py