Context isn't a prompt. It's an assembled request.
Every model call is built from pieces: instructions, what the app knows, what it looked up, the conversation, the question. In most systems the step that puts them together has no name.
When nobody owns the payload, the failures have shapes.
They repeat, they can be named, and every one of them can be reproduced.
Watch seven systems assemble, and pull a slot out →Name the parts. Fit them to a budget. Write down why.
Four planes and eleven typed slots give every piece an owner. The assembler admits what the request needs, fits it in tier order, renders it in a profile's order, and emits a trace. Same inputs, same request, every time.
Four planes. Eleven slots. Any number of items.
A plane says why a piece of context exists. A slot is a typed collection inside a plane, with its own admission rules. An item is one governed object in a slot — a policy, a retrieved chunk, a tool observation. Newcomers learn the planes; implementers use the slots; the assembler moves the items.
{{ pl.blurb }}
{{ layer.purpose }}
{{ layer.value }}
Under the hood, in raw numbers
Viability is the benchmark site: run tables, scores and the measurements behind the placement profiles. It is data heavy.
You don't need any of it to use or understand CWA. Spec, Assembler and Placement stand on their own; Viability is there if you want to look at the data yourself.
{{ viabilityCountdown }}