Beyond Single-Shot: Multi-step Tool Retrieval via Query Planning
Year: 2026
Type: preprint
Abstract: LLM agents operating over massive, dynamic tool libraries rely on effective retrieval, yet standard single-shot dense retrievers struggle with complex requests. These failures primarily stem from the disconnect between abstract user goals and technical documentation, and the limited capacity of fixed-size embeddings to model combinatorial tool compositions. To address these challenges, we propose TOOLQP, a lightweight framework that models retrieval as iterative query planning. Instead of single-shot matching, TOOLQP decomposes instructions into sub-tasks and dynamically generates queries to interact with the retriever, effectively bridging the semantic gap by targeting the specific sub-tasks required for composition. We train TOOLQP using synthetic query trajectories followed by optimization via Reinforcement Learning with Verifiable Rewards (RLVR). Experiments demonstrate that TOOLQP achieves state-of-the-art performance, exhibiting superior zero-shot generalization, robustness across diverse retrievers, and significant improvements in downstream agentic execution.
Source: arXiv (Cornell University)
Authors: Fang, Wei, Glass, James
Institution: Massachusetts Institute of Technology
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7123393885
Cites:
Cited by:
Related to: