Crate traitclaw_core 

TraitClaw Core

Core traits, types, and runtime for the TraitClaw AI Agent Framework.

This crate provides the fundamental abstractions for building AI agents:

• Provider — LLM abstraction trait
• Tool — Tool definition trait with auto JSON Schema
• Memory — 3-layer memory system (conversation, working, long-term)
• Guard — Hard boundary checks for model steering
• Hint — Contextual guidance injection
• Tracker — Runtime state monitoring
• Agent — The main agent struct with builder pattern

Quick Start

use traitclaw_core::prelude::*;

// Agent creation is done via the builder pattern
// (requires a Provider implementation)

Re-exports

pub use traits::context_manager::ContextManager;

pub use traits::execution_strategy::AdaptiveStrategy;

pub use traits::execution_strategy::ExecutionStrategy;

pub use traits::execution_strategy::ParallelStrategy;

pub use traits::execution_strategy::PendingToolCall;

pub use traits::execution_strategy::SequentialStrategy;

pub use traits::execution_strategy::ToolResult;

pub use traits::guard::Guard;

pub use traits::guard::GuardResult;

pub use traits::hint::Hint;

pub use traits::hint::HintMessage;

pub use traits::hint::HintPriority;

pub use traits::hint::InjectionPoint;

pub use traits::memory::Memory;

pub use traits::output_transformer::OutputTransformer;

pub use traits::provider::Provider;

pub use context_managers::LlmCompressor;

pub use context_managers::RuleBasedCompressor;

pub use context_managers::TieredCompressor;

pub use registries::AdaptiveRegistry;

pub use registries::DynamicRegistry;

pub use registries::GroupedRegistry;

pub use registries::TierLimits;

pub use retry::RetryConfig;

pub use retry::RetryProvider;

pub use token_counter::TikTokenCounter;

pub use token_counting::CharApproxCounter;

pub use token_counting::TokenCounter;

pub use traits::tool::ErasedTool;

pub use traits::tool::Tool;

pub use traits::tool::ToolSchema;

pub use traits::tool_registry::SimpleRegistry;

pub use traits::tool_registry::ToolRegistry;

pub use traits::tracker::Tracker;

pub use transformers::BudgetAwareTruncator;

pub use transformers::FullOutputRetriever;

pub use transformers::JsonExtractor;

pub use transformers::ProgressiveTransformer;

pub use transformers::TransformerChain;

pub use default_strategy::DefaultStrategy;

pub use traits::hook::AgentHook;

pub use traits::hook::HookAction;

pub use traits::hook::LoggingHook;

pub use traits::strategy::AgentRuntime;

pub use traits::strategy::AgentStrategy;

pub use types::action::Action;

pub use types::agent_state::AgentState;

pub use types::completion::CompletionRequest;

pub use types::completion::CompletionResponse;

pub use types::completion::Usage;

pub use types::message::Message;

pub use types::message::MessageRole;

pub use types::model_info::ModelInfo;

pub use types::model_info::ModelTier;

pub use types::stream::CompletionStream;

pub use types::stream::StreamEvent;

pub use types::tool_call::ToolCall;

pub use error::Error;

pub use error::Result;

pub use memory::compressed::CompressedMemory;

pub use memory::in_memory::InMemoryMemory;

pub use agent::Agent;

pub use agent::AgentOutput;

pub use agent::AgentOutputContent;

pub use agent::AgentSession;

pub use agent::RunUsage;

pub use agent_builder::AgentBuilder;

pub use factory::AgentFactory;

pub use pool::AgentPool;

Modules

agent

Agent — the main entry point for running AI agents.

agent_builder

Agent builder for fluent configuration.

config

Agent configuration.

context_managers

Built-in ContextManager implementations for common compression strategies.

default_strategy

Default strategy — preserves v0.1.0 agent loop behavior.

error

Error types for TraitClaw Core.

factory

Agent factory for shared-provider multi-agent creation.

memory

Memory implementations.

pool

Agent pool for managing and executing groups of agents.

prelude

Prelude module for convenient imports.

registries

Built-in ToolRegistry implementations.

retry

Retry wrapper for LLM providers with exponential backoff.

token_counter

Accurate tiktoken-based token counting.

token_counting

Token counting utilities for context management.

traits

Core trait definitions for TraitClaw.

transformers

Built-in OutputTransformer implementations for common use cases.

types

Shared types for TraitClaw.