Human-Weapon/AgentBench

Objective benchmarking and regression evaluation for AI agents, prompts, skills, models, and execution strategies.