Systematically overload an AI model with 30–40 tools to evaluate its ability to select and invoke the correct tool under heavy tool diversity.
