Locating files and functions requiring modification in large software repositories is challenging due to their scale and structural complexity. Existing LLM-based methods typically treat this as a repository-level retrieval task and rely on multiple auxiliary tools, which often overlook code execution logic and complicate model control. We propose RepoNavigator, an LLM agent equipped with a single execution-aware tool: jumping to the definition of an invoked symbol. This unified design reflects the actual flow of code execution while simplifying tool manipulation. RepoNavigator is trained end-to-end via Reinforcement Learning (RL) directly from a base pretrained model, without relying on closed-source distillation. Experiments demonstrate that RL-trained RepoNavigator achieves state-of-the-art performance, with the 7B model outperforming 14B baselines, the 14B model surpassing 32B competitors, and the 32B model exceeding closed-source models such as GPT-5 on most metrics. These results confirm that integrating a single, structurally grounded tool with RL training provides an efficient and scalable solution for repository-level issue localization.
@article{arxiv.2512.20957,
title = {One Tool Is Enough: Reinforcement Learning for Repository-Level LLM Agents},
author = {Zhaoxi Zhang and Yitong Duan and Yanzhi Zhang and Yiming Xu and Zhixiang Wang and Kun Liang and Weikang Li and Jiahui Liang and Deguo Xia and Jizhou Huang and Jiyan He and Yunfang Wu},
journal= {arXiv preprint arXiv:2512.20957},
year = {2026}
}