One Tool Is Enough: Reinforcement Learning for Repository-Level LLM Agents
This addresses the problem of repository-level issue localization for software developers, offering an efficient and scalable solution, though it is incremental as it builds on existing LLM-based methods with a novel tool integration.
The paper tackles the challenge of locating files and functions in large software repositories by proposing RepoNavigator, an LLM agent that uses a single execution-aware tool and reinforcement learning training, achieving state-of-the-art performance with smaller models outperforming larger baselines, such as a 7B model beating 14B ones and a 32B model exceeding GPT-5 on most metrics.
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.