Crate traitclaw_strategies 

TraitClaw Strategies

Built-in reasoning strategies for the TraitClaw AI agent framework.

This crate provides three reasoning strategies that implement the AgentStrategy trait:

• ReAct (react feature): Think→Act→Observe reasoning loops with tool use
• Chain-of-Thought (cot feature): Structured step-by-step reasoning
• MCTS (mcts feature): Monte Carlo Tree Search with parallel branch evaluation

All strategies are enabled by default. Use default-features = false to selectively enable only the strategies you need.

Feature Flags

Feature	Description
react	ReAct reasoning strategy (Think→Act→Observe)
cot	Chain-of-Thought reasoning strategy
mcts	Monte Carlo Tree Search strategy

Re-exports

pub use common::ThoughtStep;

pub use streaming::StrategyEvent;

pub use react::ReActStrategy;

pub use cot::ChainOfThoughtStrategy;

pub use mcts::MctsStrategy;

Modules

common

Common types shared across all reasoning strategies.

cot

Chain-of-Thought reasoning strategy.

mcts

Monte Carlo Tree Search reasoning strategy.

react

ReAct reasoning strategy (Think→Act→Observe loops).

streaming

Streaming support for reasoning strategies.