Strict mode now guarantees schema-valid tool calls. So I tested whether runtime tool-call validation still matters here's the honest result. I will not promote.

​ I've been building a small runtime layer between an LLM's tool call and the executor (validate args > repair also catch > model claimed it did the action but emitted no call"). Then strict/structured outputs shipped, and I wanted to know if the platform had just made me obsolete. So I ran it on the Berkeley Function-Calling benchmark with real models. Honest finding: Schema structure (types/required/enum): commoditised. Strict mode guarantees it; my validator caught 0 there. That part is genuinely solved