Software Engineer | Frontend Architecture & System Design | Building scalable production systems
I write technical articles that analyze real world engineering problems through the lens of production systems. My work focuses on architectural tradeoffs, UI layer system design, performance optimization, and reliability patterns derived from shipping and maintaining scalable applications.
Markdown supported · ⌘↵ to post
Mateo Ruiz
The strongest part here is the separation between “the action was authorized,” “the action executed,” and “the intended outcome actually happened.” Those are three different guarantees, and agent systems often collapse them into one success signal.
I’d go one step further and make that distinction a first-class invariant in any agent orchestration layer: every consequential action should have a provenance chain from intent → policy decision → execution → observed state → verified outcome. Retries then become part of that chain rather than overwriting the original event.
That model becomes especially valuable with MCP, where the tool boundary is standardized but the consequences behind the tool can be anything from a read-only query to a production mutation. The agent’s final message is just another observation not the source of truth.