One Tool Is Enough: Reinforcement Learning of LLM Agents for Repository-Level Code Navigation

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.

RepoNavigator helps AI better understand and navigate large software projects. Instead of relying on many complicated tools, it uses a simple way to follow how code is connected, much like how developers trace functions when reading programs. By training the system directly with reinforcement learning, the model learns to locate the parts of code that need changes more accurately and efficiently. Experiments show that this simple and focused design allows smaller models to outperform much larger and even closed-source systems on software engineering tasks.