One Tool Is Enough: Reinforcement Learning for Repository-Level LLM Agents Paper • 2512.20957 • Published Dec 24, 2025 • 1