1 paper
Tianyi Wu, Chao Ge, Peng Chen +12
AI coding agents now write much of the code that ships, yet whether that code is secure remains unclear. We introduce the Last Secure Code Benchmark, a comprehensive evaluation of…