The fifth chapter of the Use Eve track, rewritten in the Eve edition on 10
September 2026. It follows the plan agreed on 9 September and recorded in
EVE-V1-REDESIGN-TODOS.md. The receipt for the delivery is under
verification/; this file records what the chapter is for and why it is shaped
this way.
The job#
Chapter four ended with a delivered work item and a member-facing note. This chapter applies the same discipline to the content a member will actually read: Tara's editorial pipeline, Hathor's authoring drafts, Isis's quality, safety, cost and media evidence, and the launch-readiness evaluator — all reached through one read tool and five confirmed write commands.
Running example. The Enrollment guide concept, from a spark to a brief with a date on it.
Three learning promises (the cover)#
- How to ask — one workbench, one registered view, and the filters that view actually supports.
- What the numbers mean — quality, safety, cost and media evidence, each with its own population and revision.
- How a change is made — a card, the revision you reviewed, a recheck at execution, and a read-back.
Three things to remember (the close)#
- The scope is part of the answer. A rollup and a capped list are two populations.
- Every measure has a contract. Two numbers from two contracts do not compare.
- An approval names one revision. A subject that moved underneath is refused rather than overwritten.
Bridge to chapter six. The person the whole pipeline exists for.
Sections#
Four sections, opened by a section-divider that names all four.
| Section | Slides |
|---|---|
| Ask a registered view | the read contract, the Tara dossier, four supporting records, Hathor |
| Read the measures | quality, safety, cost and routing, media governance, the evaluator, an exercise |
| Act on the record | the confirmation card, sparks, the stage transition, the calendar |
| Deliver the content | the brief, the dispatch, the idea promotion, the round trip |
Treatment decisions#
- The media slide is a layer stack, not a decision tree. The plan's card grid would have flattened three questions into one list. A tree was the first choice, but three sequential questions need three question nodes and the decision-tree contract allows two — correctly, because a third nesting level does not fit at slide width. A layer stack says the same thing: three bands, with the rule that has to hold at each boundary.
- Cost and routing is a card grid, not a stat panel. The plan asked for big numbers, and the source has no numbers to show — only fields and their denominators. A stat panel would have had to invent values, so the four views keep their populations on cards instead.
- Two tables became one. Only the four Hathor views genuinely need their returned fields compared side by side; the Isis and spark tables became a card grid and a compare panel.
- The evaluator case keeps its worked numbers.
75% cleared. Still no_go.is the actual output of the pinned pure evaluator for the declared inputs, and the slide says so on its face. - No studio captures. See the receipt: the Tara pipeline and Isis quality views are admin surfaces with no component in the only tree the capture harness mounts, and a capture taken on this host would be recorded under a false environment.
What was rebound#
content/66-eve-studio-actions.jsonmerged intocontent/65-eve-studio-analysis.json; every slide id kept.catalog-source.jsonandauthoring/eve-capability-guides.json: one entry with all twenty-four slides in reading order.authoring/eve-capability-teaching.json: sixteen assignments carry the rewritten titles; every boundary was re-read and still holds.- Two layout contracts moved out of the Eve chapters, because no Eve slide uses
those layouts any more: the gated-flow contract to
scenes-cue-priorityand the relation direction contract tomember-shell.