GRAFT: Execution-Guided Genetic Recombination for Agent Design A new arXiv paper (2611.23509v1) introduces GRAFT, an execution-guided genetic recombination method for automated LLM agent design that uses parent-specific execution evidence rather than aggregate fitness alone to guide semantic crossover. Evaluated across six reasoning and programming settings with open-weight optimizer and executor models, GRAFT reached a mean test accuracy of 82.0%, beating the strongest baseline by 9.8 percentage points, with ablations holding initialization, search rules, repair policies and evaluation budgets fixed showing the execution evidence itself drove the gain. The feedback harness reuses bounded samples of successful and failed runs from fitness evaluation, withholding reference answers and requiring no additional task executions. arXiv:2611.23509v1 Announce Type: new Abstract: Automated agent design searches for executable architectures that improve LLM-based agent performance, with evolutionary methods using language-model crossover to recombine promising mechanisms for task decomposition, model coordination, and verification. This search, however, is typically guided by aggregate fitness alone, which hides the instance-level successes and failures behind each score. We introduce Execution Guided Genetic Recombination for Agent Design GRAFT , which uses aggregate fitness for selection and survival and parent-specific execution evidence to guide semantic crossover with integrated mutation. Its feedback harness reuses bounded samples of successful and failed runs from fitness evaluation, withholding reference answers and requiring no additional task executions. We evaluate GRAFT across six reasoning and programming settings using open-weight optimizer and executor models. It achieves a mean test accuracy of 82.0%, exceeding the strongest baseline by 9.8% points. Controlled ablations show that adding execution evidence improves test accuracy with initialization, search rules, repair policies, and evaluation budgets held fixed. Our results support the value of parent-specific execution evidence for evolutionary agent design.