What problem does it solve? AI-generated slide decks often lack narrative structure, visual depth, and real evidence, producing generic text-in-boxes output. This Skill manages the full deck production lifecycle — brief, narrative arc, visual composition, asset capture, build, and QA — so the final deck is commercially persuasive and self-explanatory. ## Core Features & Use Cases - Forced 8-step pipeline with gates: Brief, vibe, narrative arc, layout, compression, visual composition, build, and QA, with structured rollback routing when review fails. - Expert Mode interview: Extracts claims from source material, detects 5 gap types (case, causal, data, contrast, objection), and runs a coverage-driven interview to enrich content with real numbers and cases. - Visual composition specs: Generates per-page visual construction specs (chart types, icons, illustrative data, layout proportions) so build tools receive concrete instructions instead of raw text. - Asset pipeline and QA: Playwright screenshot capture, Pillow device mockups, concept UI skeletons, and 25+ automated QA checks including commercial scoring and geometry stability. - Use Case: Given a 30-page solution brief, initialize a 12-page deck project, run the expert interview to fill content gaps, generate visual specs, build via a PPTX/HTML build tool, and pass formal QA with a commercial scorecard. ## Quick Start Initialize a new deck project by running scripts/run_deck_pipeline.py init with your project directory, page count, and preset, then follow the gated workflow through build and QA.
