WAKE confirm-understanding

Intent gaps only. Researchable mock facts are closed below, not listed here.

1. Is Bence’s job to stand at the ring and rank, or to leave it running overnight? That picks whether skip/wake-now matter more than Up/Down.
2. When pin and briefing are both dep-safe, should pop take the first scan hit (today) or wait for Bence to pick among equals?

A. The factory is an authority path: human rank is law, questions do not guess, lowest confidence is first, pop cannot take unmet deps, implementer only works the pop. The 60s clock is the metronome. — 8/10 — gap: another cut (kiln) already made the list home; this mock made the clock hero on purpose.

B. Every station around the clock must fire on a wake. — 3/10 — gap: guesses that nobody reads before pop are costume, same family as a token in a 176-column filter.

C. Northline (overtime, bridges, 6am pin, salt, GPS briefing) is seed work to prove the loops. — 9/10 — gap: no second seed on the mock.

A. Keep the ring. The product is not “a clock with seven labels.” It is: Bence’s rank survives every later station; pop is dep-safe; salt cannot start; implementer cannot invent a second plan.

1. Best — ring with fewer wakes. Clock stays. Stations that can change the next pop keep a wake: Ideate (rank), Questions, Confidence, Roadmap, Implement. Guesses live under a question. Goals are a board the manager reads, not a separate 60s fire.
2. Also good — list as home. Same authority path, kiln-shaped: the ranked list is the hero; the clock is a header. Different cut. Do not mash with (1) on one page.

Rejected: seven costume stations + a bigger timer as the product. Clone of SparkChat feel #15/#75 four regions. A second roadmap mind. Automating the wake before rank is law.

Dumb requirement: “build a wake-clock command center with loops as stations around a 60s timer.” Costume can satisfy that without a factory.

Less dumb: Bence needs one place where his rank is law, the shakiest guess sorts first, and build cannot start on an unmet wait. The clock is how often that place moves, not what it is.

Named person: Bence. One human, one project, no CEO inbox. He must see why the next build is the next build, and see when the floor actually moved. A prettier ring that reorders his list is a miss.

The factory is for steering work, not for municipal snow. Northline is the seed: Halden overnight routes. Measured on the live mock this Project session (desktop + phone shots; add/rank, generate, promote, lowest-first 38→44→61→79→86, pop pin, salt held).

Try: Open WAKE. Clock counts. Ideate shows rank 01 overtime … 03 You added pin.

Fail today: It is fake data. Implementer % is a bar, not a worktree. Salt is honest-blocked; the rest is theatre that tells the truth about rules.

Operate: Add an idea, move rank, generate probes, nudge confidence, Pop next · dep-safe. Next-to-build must be the neighborhood pin. Salt stays unmet.

Succeed: Rank still matches what you ranked. Pop refused salt. Implementer shows the popped pin, not a second plan. Error/empty/loading still exist in the footer.

Revert: Footer → live. No bake. SparkChat untouched.

Must: lowest-first; ties on id; pop skips unmet deps.

Must not: question generator guesses; implementer starts salt; a later station silently reorders Bence.