NI, Ziyi et al. GitTaskBench: A Benchmark for Code Agents Solving Real-World Tasks Through Code Repository Leveraging. Proceedings of the AAAI Conference on Artificial Intelligence, [S. l.], v. 40, n. 38, p. 32564–32572, 2026. DOI: 10.1609/aaai.v40i38.40533. Disponível em: https://ojs.aaai.org/index.php/AAAI/article/view/40533. Acesso em: 26 may. 2026.