The map drawn from notes, before a single command — so the live trace confirms a prediction instead of discovering a path

Artifact — a hand-drawn pod-create map, assembled from your P2–P10 notes alone, that marks at each of the five area boundaries the single file you expect the request to be in — then reconciled against trace #1's spec and corrected in red where the two disagree. Where your map is wrong is where a phase did not land, and that red mark is the real output of this exercise: it names, before you spend any RAM, which seam you are about to fail to cite.

Rests on — the whole descent. This is the one phase prerequisite on all of P2–P10; an area you did not enter is a seam you cannot draw, and drawing it wrong here is cheaper than discovering it wrong live.

Topology — none. This is paper and source only; pair is not up yet and there is nothing to cost.

Read — the module framing and trace #1's spec, but only after you have drawn the map from memory. Reading the spec first defeats the exercise: the point is the delta between what eight phases left in your notes and what the source actually does.

Question to answer from the source: the trace crosses five areas. At each of the four internal seams, which single function hands the object to the next area — the last frame in area N, the first frame in area N+1? Name both, from your notes, before you open the spec.

Build — from your P2–P10 notes alone, draw the full path of kubectl run nginx and mark, at each area boundary, the file you expect the request to be in. Then open the spec and reconcile. Correct every wrong guess in red. Live-verify nothing yet — the live modules below do that, seam by seam; this map is the prediction they test.

Verify from outside — the map is honest when every arrow carries a predicted file name (right or wrong), not a subsystem label. "The apiserver handles it" is not a prediction; handlers/create.go is. A reader should be able to hold your map beside the spec and count the red corrections.

Expect — corrections concentrated at the seams whose phase you rushed. A clean map is either mastery or a lie you are about to pay for at the seam it hides; expect at least one red mark, and treat a wholly clean map as the thing to distrust.

Write down — the predicted map with a file per seam, kept beside you and corrected in red as Seam A, Seam B and Seam C prove or disprove each guess. It is one of the phase's committed written artifacts.

Footprint note — none. No cluster, no build guest; the only cost is your own paper. The ceiling is untouched until Seam A brings pair up.

Teardown — nothing was created but your notes, which you keep. No topology to release.