We are missing some building blocks IMO. We need a good abstraction for defining the invariants in the structure of a project and communicating them to an agent. Even if we had this, if a project doesn’t already consistently apply those patterns the agent can be confused or misapply something (or maybe it’s mad about “do as I say not as I do”).
replyI expend a lot of effort preparing instructions in order to steer agents in this way, it’s annoying actually. Think Deep Wiki-style enumeration of how things work, like C4 Diagrams for agents.
Yes, great points.
replyAgentic workflows can mix algorithmic + agentic steps. There's a design pattern we call "DataOps" which is all about this - algorithmic extraction then an agentic step delivering a safe output.