![[2602.20426] Learning to Rewrite Tool Descriptions for Reliable LLM-Agent Tool Use](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.20426] Learning to Rewrite Tool Descriptions for Reliable LLM-Agent Tool Use
Summary
This paper presents Trace-Free+, a curriculum learning framework designed to enhance the quality of tool interfaces for LLM-based agents, improving their performance in diverse settings.
Why It Matters
As LLM-based agents increasingly rely on tool interfaces, optimizing these interfaces is crucial for scalability and generalization. This research addresses the limitations of existing methods by proposing a novel framework that enhances agent performance without requiring extensive execution traces, making it applicable in privacy-constrained environments.
Key Takeaways
- Trace-Free+ framework improves tool interface quality for LLM agents.
- Supports generalization to unseen tools, enhancing scalability.
- Addresses limitations of existing methods reliant on execution traces.
Computer Science > Artificial Intelligence arXiv:2602.20426 (cs) [Submitted on 23 Feb 2026] Title:Learning to Rewrite Tool Descriptions for Reliable LLM-Agent Tool Use Authors:Ruocheng Guo, Kaiwen Dong, Xiang Gao, Kamalika Das View a PDF of the paper titled Learning to Rewrite Tool Descriptions for Reliable LLM-Agent Tool Use, by Ruocheng Guo and 3 other authors View PDF HTML (experimental) Abstract:The performance of LLM-based agents depends not only on the agent itself but also on the quality of the tool interfaces it consumes. While prior work has focused heavily on agent fine-tuning, tool interfaces-including natural language descriptions and parameter schemas-remain largely human-oriented and often become a bottleneck, especially when agents must select from large candidate tool sets. Existing approaches to improving tool interfaces rely on execution traces, which are frequently unavailable in cold-start or privacy-constrained settings, and typically optimize each tool independently, limiting scalability and generalization to unseen tools. We propose Trace-Free+, a curriculum learning framework that progressively transfers supervision from trace-rich settings to trace-free deployment, encouraging the model to abstract reusable interface-usage patterns and tool usage outcomes. To support this approach, we construct a large-scale dataset of high-quality tool interfaces using a structured workflow over a diverse collection of tools. Experiments on StableToolBench and Res...
