How I build the harness behind the practice. Real work, shipped to production.

Why Chat Models Keep Reframing Everything

May 2026 — A sourced, curious look at why chat models so often reach for the same negation-then-reframe sentence.

The First Screen Is a Control Surface

May 2026 — How Gauges Green turned agent startup from a context dump into a compact operating instrument.

The Agent Was Fine. The Delivery Surface Was Dark

Apr 2026 — An agent can generate the right artifact and still fail if the delivery surface has not been paired, probed, and observed independently.

I Built an AI Design Critic That Judges My Apps Against Apple's HIG

Feb 2026 — Code analysis, rendered screenshots, and a full-app crawler enforce design standards from commit to deploy.

Your AI Agent Doesn't Know What to Do Next

Feb 2026 — The harder operational question is what the operating harness should surface first.

The CI Pipeline I Actually Want

Feb 2026 — Two stages across iOS, web, backend, and database, informed by Fowler, Beck, and DORA.

How I Made Claude Code Do Its Own Deployments

Feb 2026 — The useful pattern is the operating harness around deploy work: permission, boundaries, verification, and recovery.

Compiled Context: How I Stopped Fighting Agent Memory and Started Engineering It

Feb 2026 — A 350-line Python script compiles source-backed operating context for each agent session.

My AI Controls My Webcam

Feb 2026 — It captured a frame, analyzed the lighting, adjusted the hardware settings, and iterated. No app, no sliders, just a conversation.