Refresh the Impeccable product experience

Rework the landing page proof, steering demo, feature grid, slop catalog, detector coverage, theming, Live workflow, and responsive behavior.\n\nAI-assisted implementation by OpenAI Codex.

skill/scripts/concept-seed.mjs

@@ -1,6 +1,6 @@
 #!/usr/bin/env node
 /**
- * Concept-seed picker: the dice half of the new-work concept procedure.
+ * Surface-concept seed: the dice half of new-work's task-composition procedure.
  *
  * The model derives a grounded shortlist of candidate FORMS from the
  * audience's world and the subject's cultural home (see
@@ -12,10 +12,9 @@
  *
  * This script rolls them from outside, the same trick that made the
  * palette seed work:
- *   - BUILD INDEX (2-5): which entry of the model's own resonance-ordered
- *     shortlist to build. The dice never choose an ungrounded ingredient;
- *     they only refuse the argmax rut. (Index 1 is excluded: that's the
- *     concept every run would ship anyway.)
+ *   - PROMOTED INDEX: which entry of the model's own resonance-ordered
+ *     shortlist must be taken seriously beside its favorites. The dice never
+ *     choose an ungrounded ingredient; they only refuse the argmax rut.
  *   - CHALLENGERS (3): outside forms from concept-ingredients.json, weighed
  *     against the derived candidates on exactly two axes — audience
  *     identification and product clarity. They win only when they beat the
@@ -68,21 +67,23 @@ for (let i = 0; picks.length < 3 && i < 60; i++) {
 }
 
 process.stdout.write(`CONCEPT SEED (key: ${key}; rerun with --from ${key} to reproduce this roll)
-BUILD INDEX: ${buildIndex}
-  After ordering your derived candidates by resonance, build the page whose
-  form comes from candidate number ${buildIndex}, exactly as if it had ranked
-  first: full commitment. Your top-ranked candidate is what every run in this
-  category would ship; the assignment exists to refuse that rut, not to
-  punish it.
+PROMOTED INDEX: ${buildIndex}
+  After ordering the task's grounded structural candidates by resonance,
+  promote candidate ${buildIndex} into the serious shortlist. In an attended
+  run, present it beside the strongest materially different candidates and
+  let the user select or revise the surface concept. In a truly unattended
+  run, use it when it survives audience identification and product clarity.
+  The promotion exists to refuse the model's ranking rut, not to outrank the
+  user or the brief.
 CHALLENGERS (weigh against your derived candidates on the same two axes,
 audience identification and product clarity; a challenger wins only when
 it beats the grounded list on both):
   1. ${picks[0]}
   2. ${picks[1]}
   3. ${picks[2]}
-If a challenger wins, it replaces the assigned candidate. If the surface is
-an existing world whose incumbent carries a deliberate, ownable idea, the
-incumbent IS the chosen candidate: intensify its lineage and ignore the
-roll entirely. The same override applies when the user, PRODUCT.md, or
-DESIGN.md pins a direction: pinned direction beats the roll, always.
+If a challenger survives, it may enter the shortlist as a structural option.
+PRODUCT.md and DESIGN.md constrain every candidate's identity vocabulary;
+they do not cancel task-level composition. A user- or brief-pinned surface
+concept beats the roll, always. The seed never authorizes a new palette,
+type system, material world, or unfamiliar control behavior.
 `);