Search papers, labs, and topics across Lattice.
This paper introduces Tool Primitives, a novel approach that utilizes natural language as an interface for tool invocation, addressing the challenges of brittle reasoning and performance degradation in large tool catalogs. By integrating a centralized repository called ToolFace, which houses over 25,000 functions, the framework allows LLMs to dynamically retrieve relevant tools without the need for explicit API schema enumeration. The proposed HEART framework enhances this system by providing a Planner, Router, and Verifier for efficient multi-step execution and feedback-driven recovery, resulting in a 10% performance improvement over SFT-based models and a significant reduction in API costs.
HEART achieves 84% task completion on real-world tasks, outperforming leading models while slashing API costs by up to 85%.
Large language models (LLMs) augmented with external tools have demonstrated remarkable capability in solving complex real-world tasks. However, existing approaches suffer from two key challenges: brittle multi-step and multi-turn reasoning caused by incompatible tool output types and API schemas, and performance degradation under large tool catalogues. To address these, we introduce \textbf{Tool Primitives}, a design that replaces rigid API schema-based invocation with natural language as the interface for tool calling, where each tool is wrapped with an LLM interface that handles schema resolution and execution internally, enabling natural inter-tool communication for nested and multi-turn tool calling. Building on Tool Primitives, we host \textbf{ToolFace}, a centralized repository of 25,519 functions from which LLMs dynamically retrieve only the relevant tools at inference time, eliminating the need to enumerate raw API schemas in context. To orchestrate Tool Primitives and ToolFace reliably in complex settings, we further propose \textbf{HEART}, a \textbf{H}arness \textbf{E}ngineering framework via \textbf{A}gent-native, \textbf{R}eusable \textbf{T}ool Primitives, comprising a Planner, Router, and Verifier that jointly support dynamic tool invocation planning, multi-step execution, and feedback-driven recovery. Experiments on five benchmarks demonstrate that HEART outperforms SFT-based models by $10\%$ on average and surpasses GPT-5.4, Claude-4.6-Sonnet, and Gemini-3.1-Pro by $6\%$ on average while reducing API cost by up to $85\%$. On 50 real-world tasks, HEART achieves $84\%$ task completion, $3.8\times$ the average of three frontier commercial models ($22\%$).
