deck-craft

Self-contained slide-deck skill spanning three output targets from one body of deck craft: self-contained fixed-stage HTML presentations, lecturn.deck/1 JSON (and .pptx through the deckconv round-trip), and Diolog investor decks assembled from the 140-template / 12-recipe slide library. Carries its own visual-design layer rather than depending on a general design skill: consuming a supplied DESIGN.md or token file as binding context, authoring a direction when no brand exists (scheme / formality / density, subject-mined palettes, the signature element, the swap test), typography sized for projection with tracking and measure rules, colour strategy and accent budgets, five-vector hierarchy and cross-slide rhythm, anti-slop rules, and an accessibility floor calibrated for projection. Deck-specific narrative discipline: the title sequence as the deck's argument, speaking-vs-reading density, one idea per slide, deliver-the-whole-count, and the linear-flow usability rules (per-slide trunk test, recognition over recall, the honesty gate on charts and claims). Ships a per-slide build gate and a delivery review built on the verification epistemics that rendering an image is not seeing one, that the question you bring to a capture determines what you see in it, and that a clean gate means no known defect rather than verified. Direction evidence is read progressively from the template libraries on the machine (frontend-slides bold pack, open-design deck templates, portable design systems) without bulk-reading them. v1.1 folds the slide-template library into the skill: 200 layouts across 27 families (the established 140 verbatim plus 60 additions from an ADHD divergent-ideation gap analysis covering decision/ask, provenance, uncertainty, orientation, objection, flow, product/engineering, customer/commercial and people/plan — the three structural holes being that no template in the original 140 renders an ask, that the library is listed-company-shaped, and that nothing carries provenance or uncertainty as structure), 21 recipes (12 investor plus 9 business/product/engineering spines), the full slot contract and every validator rule including three deck-level floors that previously existed only in validate-deck.mjs (a headline figure authored as text at >=40px must be a stat, four or more slides on one ground fails, fewer than four distinct font sizes fails), a 34-system visual direction index with its progressive read rule, and layout-specs.md — a per-template block for all 200 giving zones fitted to the 1280x720 frame, the focal point and the hierarchy vectors that earn it, type roles, accent placement, the UX job, the failure mode, and measured caps that override the catalogue's authored-from-intent max values. v1.2 closes the pipeline-integration gaps: the mandatory render_deck loop (stage, render every slide through the same adapter and renderer the apply path uses, fix every blocker, re-render) as the deck producer's visual feedback rail, the themeTokens passthrough that makes the preview the deck that actually ships, picking sizes from the theme's derived typography.scale ladder instead of inventing a number per element, the fontSize field name, and the job-envelope contract (USER_REQUEST.title verbatim, emit only stage:planning, never emit structure or content events). v1.3.0 folds in the portable ideas from the impeccable skill. Direction-finding gains an anti-convergence procedure in visual-craft.md: name the occasion's rut AND its predictable opposite before shortlisting (board pack to navy-and-grey, and board-pack-but-not-navy to near-black-with-one-accent, are the same reflex one tier apart), derive candidates from the audience's own world including its graphic and screen traditions spanning at least three material families, and don't take your top-ranked candidate by default because it is what every run on the brief produces; direction-index.md becomes evidence for the shortlist rather than the thing that generates the candidates; the swap test extends one tier deeper to the second-order reflex. Every direction round now carries a standing exit, the convention played straight, always offered and never recommended, executed at full fidelity when taken, because board and regulatory audiences legitimately want the category standard and should not have to argue past the anti-slop rules to get it. The chosen direction is written into the artifact as a five-block contract (THESIS, OWN-WORLD, STORY, COVER, FORM) as an HTML comment or x.diolog.direction, since a direction declared only in conversation drifts by slide 9, and deck-review.md's delivery pass now audits the finished deck against those blocks promise by promise with mandatory TYPE and MATERIAL rows classifying every committed element match, adaptation, missing or contradicted. Reviews close with a computed disposition (ship, fix or rebuild) reported verbatim, escalate to a rebuild directive naming slides to re-derive and assets to produce when the direction is contradicted across the deck rather than on a slide or two (a patch list against a deck that failed wholesale launders the rejection into an approval), and add a verdict pass scoring each fix resolved, partial or unresolved against the recapture, because an account of what was fixed is not evidence any more than a screenshot nobody opened is. html-deck.md gains a medium gate ahead of its treatment rules: a photograph, a site, a figure or a texture named as one is raster whatever the stack and layered CSS gradients are not a texture medium, while rules, hairlines, drawn marks, flat shape systems and diagrams are authored SVG or CSS, plus a written quantity commitment for fields and textures and a rule to judge the composite rather than the asset. visual-craft.md's anti-slop section adds imitation material (CSS bevels, faux letterpress, fake foil, stamped-metal and chalk effects standing in for a material the slide never renders, the most reliable machine-made tell and one decks reach for on covers), zero-offset halos and hard offset block shadows, and an explicit carve-out that the SVG ban targets pictures and never geometry. The first slide is now gated hardest, with scale, density and material judged as quantities rather than impressions.