JTPRO: A Joint Tool-Prompt Reflective Optimization Framework for Language Agents

TL;DR

JTPRO is a framework that enhances language agents' ability to accurately select and utilize tools by iteratively optimizing prompts and tool schemas, leading to improved reliability in multi-tool environments.

Contribution

It introduces a joint optimization method that refines instructions and tool schemas to improve tool selection and argument filling in language agents.

Findings

JTPRO outperforms baselines by 5%-20% on success rate.

Joint optimization of instructions and schemas is more effective than isolated tuning.

JTPRO improves tool-calling reliability in multi-tool benchmarks.

Abstract

Large language model (LLM) agents augmented with external tools often struggle as number of tools grow large and become domain-specific. In such settings, ambiguous tool descriptions and under-specified agent instructions frequently lead to tool mis-selection and incorrect slot/value instantiation. We hypothesize that this is due to two root causes: generic, one-size-fits-all prompts that ignore tool-specific nuances, and underspecified tool schemas that lack clear guidance on when and how to use each tool and how to format its parameters. We introduce Joint Tool-Prompt Reflective Optimization (JTPRO), a framework for improving tool-calling reliability in trace-supervised settings by iteratively using rollout-driven reflection to co-optimize global instructions and per-tool schema/argument descriptions for accurate tool selection and argument instantiation in large tool inventories. JTPRO is designed to preserve only tool-local cues needed for correct disambiguation and slot filling. We evaluate JTPRO across multi-tool benchmarks, which account for different number of tools using three metrics: Tool Selection Accuracy (TSA), Slot Filling Accuracy(SFA), and Overall Success Rate(OSR) (correct tool + correct slots + correct values). JTPRO consistently outperforms strong baselines, including CoT-style agents, and reflective prompt optimizers such as GEPA by 5%-20% (relative) on OSR. Ablations show that joint optimization of instructions and tool schemas is more effective and robust than optimizing either component in isolation.