Olinave
Free Clinic · live online

Why your coding agent ignores you — live, in one free hour.

A free, live hour: the controls that decide whether Claude Code follows your rules, the warning signs when one isn't holding, and a single fix demoed live.

The no-commitment front door to the Control Surfaces framework — a demo-driven teaching hour, then open Q&A.

Register on Luma → Free · next session July 16
What it is

Think of Claude Code like a dashboard.

Your agent has controls that decide whether it follows your rules — your Control Surfaces. Most people are driving without knowing the dashboard is there.

The Clinic shows you the controls and how to use them. We anchor on the one surface everyone knows — CLAUDE.md — place it in the full map, then go deep on the three failure modes most visible from there.

The warning lights

Three ways a rule gets ignored.

Three of the twelve named failure modes — the three most visible from CLAUDE.md:

  • Compliance Illusion

    A passing result doesn't prove your rule was followed — the agent can land the right output while never consulting the instruction.

  • Critical Rules in Prose

    A zero-tolerance “MUST” / “NEVER” written in a sentence reads like a hard constraint. Structurally, it only suggests.

  • Context / Attention Dilution

    The rule doesn't fail because it's absent — it fails because everything else in the window is competing for the agent's attention.

What you leave with

One complete, proven ‘aha’.

01

Why it happens

Why the agent walks past your CLAUDE.md — named, not hand-waved.

02

What it looks like

How to spot the warning signs in a real setup, live on screen.

03

The one move

We take a prose rule and make it actually hold — a surface that gates instead of asks.

Philip Forshaw
Who runs it

Philip Forshaw

Olinave · ex-Apple AI Strategy & Operations

Ex-Apple AI Strategy & Operations lead, with decades of training and delivery behind him. The framework is patent-pending.

The method is a named taxonomy of the recurring failure modes, grounded in controlled experiments and Anthropic's own published guidance — then measured across 359 analysed Claude Code setups. Not theory; the hour works from the evidence.

Take the free hour

Save your seat.

Free, live, ~60 minutes. Can't make July 16? It runs again July 30. Want the whole framework end to end? It ladders into the Masterclass.

Register on Luma →

Free · live online · 9:00 AM PDT