Search for a command to run...
SwS: Self-aware Weakness-driven Problem Synthesis in Reinforcement Learning for LLM Reasoning