All software design assumes an actor with intent. The well-intentioned actor and the malicious actor are archetypes at opposite ends of an intent spectrum. Every design paradigm — the pit of success, the fortress — is a response to where the designer places the actor on that spectrum. The spectrum’s organizing axis is intent itself.
This binary held for decades. But it breaks when you introduce an actor that has no intent at all: the AI coding agent. Placing it on the intent spectrum is a category error — the entire axis is organized around something the agent doesn’t have.