The paper introduces EdgeGen, a framework that automatically generates synthetic edge-case tasks by extracting compliance rules from agent specifications and building database-grounded scenarios designed to violate them, enabling annotation-free finetuning and agent-harness optimization. On the tau2bench airline domain, finetuned models show 2-42 percentage-point gains and harness optimization yields 10-30 percentage-point improvements for Gemma-4-e4b.
